BY JIM BURATTI AND TIM SIMONSON

For generations, people in North America have been able to place and receive highly reliable voice calls on a legacy circuit-switch network. It is so commonplace, we don't even think about it until there is a problem, which is very rare. Behind each phone call is a rugged infrastructure that carries this traffic. One of the hallmarks of this legacy system infrastructure is a well thought-out power availability system. The highly reliable level of service and availability in the legacy world are the comparative standard to which convergent networks (VoIP and packet voice networks) will be held to. Therefore, as the age of network convergence accelerates, it is critical that network architects, engineers, operations teams, and even end users create an equally reliable and available power environment.

Without an end-to-end power availability plan, there is little chance of a packet-based network meeting the user's expectation for reliability in voice, fax, or video applications. However, much like a convergent network is designed to be more efficient (from a total cost of ownership perspective), scalable, and manageable, the power architecture of new networks can be equally improved. For example, some companies' routers and remote servers can be remotely managed -- so too can UPSs using Simple Network Management Protocol (SNMP), IP interfaces, contact closures, or out-of-band dial-in modems to extend control from the Network Operations Center (NOC) to any site around the globe in real time.

ADDING UP THE OPTIONS
There are several differences and advantages to designing an AC powered network.

One of the first differences between the legacy and convergent networks is the power environment itself. The phone companies run their equipment on -48vDC (minus 48 volts direct current - versus AC alternating current) power in a central office environment. New Age carriers often operate their equipment on AC power or a combination of AC power and DC power. When designing a distributed, flexible, manageable, and reliable network, it is often easier to standardize on AC power. AC power utility service is readily available, therefore all that is needed is an intelligent UPS that can condition the power and provide the desired battery runtime to the equipment. Contrast this with DC power, which takes in AC power and then uses an expensive system to invert the existing power down to -48vDC. No question it can be a reliable architecture, but it can be overkill in terms of total cost of the infrastructure.

AC UPSs are fully manageable just like any other network device. In other words, using SNMP, IP interface, or dial-in out-of-band management via an independent POTS (plain old telephone service) modem, a UPS can be tied into the global NOC. By making the UPS part of the core network infrastructure, users can increase the manageability and reliability of their systems with real-time access to data which reports the power status at every node or Point of Presence (PoP) on the network. This is a significant improvement relative to the rudimentary dry contact closure monitoring or basic on-site diagnostics provided by the average DC power infrastructure.

For example, leading UPS manufacturers offer interfaces with network device monitoring packages such as CA Unicenter, Tivoli, HP OpenView, etc. Or, the interfaces can be equally managed by using a proprietary system and the UPS vendor's own software tools.

The return on investment and operational cost savings of a fully monitored and managed power infrastructure pay for themselves through the ability to dispatch planned maintenance calls rather than emergency personnel to address or diagnose an issue. For example, UPS batteries will need to be replaced on average between three and five years after installation. By using intelligent monitoring and management, users can identify the battery health and status of each UPS unit from the NOC, and then plan accordingly. This is a practical example of proactively managing the cost of ownership. Without such monitoring, the alternative is on-site service inspection and testing or a shot-in-the-dark strategy of replacing all batteries at the same time with no regard to whether they actually need to be replaced.

Another smart power management opportunity is a power outlet control device deployed in conjunction with a UPS. This type of device allows the NOC to reboot hung network equipment without dispatching personnel to a site. The control and reboot function can be done over the network, or if the network itself is hung, via a regular telephone line into the device.

RAPID LINK STAYS UP
A good example of a company that has taken a comprehensive approach to their power infrastructure is Atlanta-based Rapid Link, Inc.. Rapid Link is an Internet-based communications service provider that routes voice, data, and fax traffic over switched and IP networks. Rapid Link has the ability to expand beyond voice into new Internet protocol, data, and e-commerce applications, while continuing to effectively control transmission costs. Each of Rapid Link's PoPs is fully redundant by operating a UPS with generator backup and restricted, secure access.

From a power perspective, Rapid Link has deployed UPS units globally to their network node locations in conjunction with a management card that gives users the ability to manage each device around the globe into their Atlanta NOC. This system means they can install PoPs in almost any location very quickly with eight hours of runtime, and full management and monitoring capability from their fully staffed NOC.

When the UPSs perform periodic scheduled battery testing, the company's operations staff receives data on all units that may need servicing in the near future. This advance warning notice allows them to probe further by pinpointing the device or scheduling a PoP or data center visit by the appropriate technician, according to Dr. Bill Kennedy, chief technology officer with Rapid Link.

"Without this intelligent communication with the UPS, we would often have no other choice than to wait for a unit to fail and then create a trouble ticket to correct the situation," Kennedy added. Rapid Link realizes that it is very important that both a facilities and IT approach to powering the network are considered when configuring the right application for convergent carriers and legacy operators.

Rapid Link was able to standardize on a UPS model that met their goals and was a widely available plug-and-play system. This matched their objectives for a speedy network rollout without intensive on-site installation and engineering required. The reliability of the systems has translated into optimized network operations, which means optimized revenue for Rapid Link's global voice network. The reliability and high customer satisfaction experiences translate into word-of-mouth recommendations and new customers for Rapid Link. Every time a call is made on their network, its business reputation is tested. Because of a sound power protection strategy, the company has insured that inevitable power anomalies are not going to jeopardize its operations.

The challenge of convergent carriers to achieve their network availability goals must include a sound power strategy. Leveraging the flexibility, manageability, scalability, and cost of AC UPSs can actually contribute to attaining aggressive business and operational goals. c

Jim Buratti is Internet market specialist for American Power Conversion (APC), and may bee-mailed at [email protected]. Tim Simonson is manager, service provider customer segment, for APC. APC is a global leader in the designing, manufacturing, and marketing of power protection equipment, including surge suppressors, uninterruptible power supplies (UPS), power conditioning equipment, and related software for computer and computer-related equipment. Rapid Link has deployed the APC Matrix UPS units globally, in conjunction with the APC Web/SNMP management card. For more information about APC, visit the company's Web site at www.apcc.com.

BY CHRIS HOLDER

As carriers speed to deploy integrated voice, data, and video communications, additional equipment units are being rapidly installed to build out new arms of an underlying infrastructure. The enabling infrastructure that supports these new communications services is a complex array of data communications and telecom equipment. A critical success factor of implementing integrated communications is ensuring that networks maintain maximum availability - leaving zero tolerance for downtime due to crashed or failing equipment units.

CLECs, ILECs, IXCs, and ISPs are building new business models and service offerings based on the dependability of data communications and telecom internetworking equipment. While these carriers rely on battery power (DC) or power conditioning/UPS (AC) to ensure power availability, a dependable power supply alone cannot be relied on to maintain the operational status of these devices. Despite high reliability assurances from hardware manufacturers and software programs, equipment units do malfunction. When equipment units "lock up" or fail, the most proven method to restore the device to its operational state is to cycle the power - a reboot. However, when equipment units are remotely located in a distant POP site, co-location facility, or telco central office, performing the reboot can be a costly function of time, distance, and third-party expense. At risk for the service provider are lost revenue, customer dissatisfaction, decreased productivity, and potential service level agreement (SLA) penalties.

Consider what happens when a remote router locks up. The router's power cord is connected to one of the UPS's multiple power outlets, but a UPS does not have the ability to power cycle an individual power output receptacle. The recovery action choices available to the network control center are limited. The first option is a sledgehammer approach whereby an operator commands the UPS to simultaneously power cycle the UPS and all its attached devices. The second option is to find a technician that can be dispatched to the remote location to power cycle only the locked-up router. Neither choice is attractive or efficient.

POWERING AN INFRASTRUCTURE
One leading Internet telephony service provider has found a solution for not only addressing the problem of crashed internetworking devices in remote locations, but also faster installation and build-out of its infrastructure. Net2Phone, the first company to bridge the Internet with telephone networks, is using remote power managers to ensure its global network of PC-to-phone, phone-to-phone, fax-to-fax, and e-commerce solutions for businesses and consumers maintains maximum uptime and availability.

Net2Phone offers the most advanced telecommunications application harnessing the global power of the Internet, and has enabled more than 1,000,000 customers to place phone calls over the Internet at rates up to 95 percent less than current phone rates. Creating this new Internet telephony service has required Net2Phone to build a network consisting of servers, routers, CSU/DSUs, and other equipment units placed at several international POPs. With a remote power management unit in place, the recovery of a locked-up equipment unit is easy. A remote power manager provides a logical, software-controlled interface to individual power outlet ports. Now, upon receiving notification of a locked-up equipment unit, a Net2Phone network operations center (NOC) can immediately reach out to power cycle the individual equipment unit and quickly return it to an operational status -- without interrupting all the equipment attached to the UPS.

In addition to remotely rebooting devices, Net2Phone also utilizes a unique feature of the power managers that it has deployed at more than 40 international POP sites. Asynchronous communications ports are built into each unit, which allow a network operator to establish serial communications to the management port for all the internetworking devices located in an equipment rack.

Net2Phone found that being able to remotely establish a communication session to the management console port on an internetworking device allows the service provider to more quickly build its global infrastructure. As Net2Phone deploys servers and other internetworking devices, the service provider distributes the devices from a central "staging" management center. The units are shipped to international locations, where technicians follow master cabling and installation instructions from the management center. Upon successful installation of the equipment units, the Net2Phone management center connects to the power manager to "power on" the devices. Then, via the asynchronous communications ports, Net2Phone establishes a serial communications session with each of the internetworking devices to remotely configure and monitor each device in its Internet telephony POP site. This process allows Net2Phone to dramatically reduce the time it takes to add new international POPs to its global infrastructure.

Chris Holder is product marketing manager for Server Technology. Server works with companies like Net2Phone, which uses the company's Sentry Remote Power Managers. The Sentry Remote Power Managers are an industrial product line that enables remote power control (off, on, reboot, graceful shutdown) of servers and internetworking devices in remote equipment rooms, POP sites, telco central offices, and other co-location facilities. For more information, visit the company's Web site at www.servertech.com.